HiTZ/mt-hitz-en-eu

Hitz Center’s English-Basque machine translation model

Model description

This model was trained from scratch using Marian NMT on a combination of English-Basque datasets totalling 20,523,431 sentence pairs. 9,033,998 sentence pairs were parallel data collected from the web while the remaining 11,489,433 sentence pairs were parallel synthetic data created using the Google Translate translator. The model was evaluated on the Flores, TaCon and NTREX evaluation datasets.

• Developed by: HiTZ Research Center & IXA Research group (University of the Basque Country UPV/EHU)
• Model type: traslation
• Source Language: English
• Target Language: Basque
• License: apache-2.0

Intended uses and limitations

You can use this model for machine translation from English to Basque.

At the time of submission, no measures have been taken to estimate the bias and toxicity embedded in the model. However, we are aware that our models may be biased since the corpora have been collected using crawling techniques on multiple web sources.

How to Get Started with the Model

Use the code below to get started with the model.

from transformers import MarianMTModel, MarianTokenizer
from transformers import AutoTokenizer
from transformers import AutoModelForSeq2SeqLM

src_text = ["this is a test"]

model_name = "HiTZ/mt-hitz-en-eu"
tokenizer = MarianTokenizer.from_pretrained(model_name)

model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=T
rue))
print([tokenizer.decode(t, skip_special_tokens=True) for t in translated])`

The recommended environments include the following transfomer versions: 4.12.3 , 4.15.0 , 4.26.1

Training Details

Training Data

The English-Basque data collected from the web was a combination of the following datasets:

Dataset	Sentences before cleaning
CCMatrix v1	7,788,871
EhuHac	585,210
Ehuskaratuak	482,259
Ehuskaratuak	482,259
Elhuyar	1,176,529
HPLT	4,546,563
OpenSubtitles	805,780
PaCO_2012	109,524
PaCO_2013	48,892
WikiMatrix	119,480
Total	15,653,108

The 11,489,433 sentence pairs of synthetic parallel data were created by translating a compendium of ES-EU parallel corpora into English using the ES-EN translator from Google Translate.

Training Procedure

Preprocessing

After concatenation, all datasets are cleaned and deduplicated using bifixer (Ramírez-Sánchez et al., 2020) for identifying repetions and cleaning encoding problems and LaBSE embeddings to filter missaligned sentences. Any sentence pairs with a LaBSE similarity score of less than 0.5 is removed. The filtered corpus is composed of 9,033,998 parallel sentences.

Tokenization

All data is tokenized using sentencepiece, with a 32,000 token sentencepiece model learned from the combination of all filtered training data. This model is included.

Evaluation

Variable and metrics

We use the BLEU and TER scores for evaluation on test sets: Flores-200, TaCon and NTREX

Evaluation results

Below are the evaluation results on the machine translation from English to Basque compared to Google Translate and NLLB 200 3.3B:

####BLEU scores

Test set	Google Translate	NLLB 3.3	mt-hitz-en-eu
Flores 200 devtest	20.5	13.3	19.2
TaCON	12.1	9.4	8.8
NTREX	15.7	8.0	14.5
Average	16.1	10.2	14.2

####TER scores

Test set	Google Translate	NLLB 3.3	mt-hitz-en-eu
Flores 200 devtest	59.5	70.4	65.0
TaCON	69.5	75.3	76.8
NTREX	65.8	81.6	66.7
Average	64.9	75.8	68.2

Additional information

Author

HiTZ Research Center & IXA Research group (University of the Basque Country UPV/EHU)

Contact information

For further information, send an email to [email protected]

Licensing information

This work is licensed under a Apache License, Version 2.0

Funding

This work is funded by the Ministerio para la Transformación Digital y de la Función Pública - Funded by EU – NextGenerationEU within the framework of the project ILENIA with reference 2022/TL22/00215337, 2022/TL22/00215336, 2022/TL22/00215335 y 2022/TL22/00215334

Disclaimer

Click to expand

The models published in this repository are intended for a generalist purpose and are available to third parties. These models may have bias and/or any other undesirable distortions. When third parties, deploy or provide systems and/or services to other parties using any of these models (or using systems based on these models) or become users of the models, they should note that it is their responsibility to mitigate the risks arising from their use and, in any event, to comply with applicable regulations, including regulations regarding the use of Artificial Intelligence. In no event shall the owner and creator of the models (HiTZ Research Center) be liable for any results arising from the use made by third parties of these models.